Available online at www.sciencedirect.

com

ScienceDirect
Procedia Economics and Finance 32 (2015) 1827 – 1836

Emerging Markets Queries in Finance and Business

Detecting and Preventing Fraud with Data Analytics

Adrian Bănărescua,b,*
a
Institutul de Economie Nationala, Calea 13 Septembrie nr.13, Bucharest 050711, Romania
b
Agenаia Naаională de Administrare Fiscală, Splaiul Independenаei, 202A, sector 6, Bucharest 060022, Romania

Abstract

Although fraud is not a new issue, the current financial crisis has enlightened that fraud occurs mainly during a recession,
as compared with normal periods of economic growth. In counterbalance to the slow economic recovery, managers need to
start a series of antifraud measures, as a leverage of cost control, while reducing available resources. Fraud involves
inclusively significant financial risks which may threaten profitability, and the image of an economic entity. In these
circumstances, in which development of the IT systems plays a central role in the creation of competitive companies, the
amount of processed data has grown exponentially. Internal control team members should need to look at every transaction
that takes place, but, unfortunately this issue can no longer be manually performed, requiring the use of data analysis tools
and programs. Since the companies usually operate with large volumes of data, it is absolutely necessary to implement such
processes of continuous monitoring, in order to identify anomalies in the data stream or behavioral patterns, potentially
fraudulent. Such new and significant information will be later used in directing investigations, as well as to make
recommendations to improve the control activities. We strive to provide an overview of the way in which technology can
be implemented to improve fraud prevention and detection, inside of a public or private economic entity.

©©2015
2015 The
Authors. Published
Authors. by Elsevier
Published B.V. This
by Elsevier B.V. isThis
an open
is an access article article
open access under the CC the
under BY-NC-ND license license
CC BY-NC-ND
(http://creativecommons.org/licenses/by-nc-nd/3.0/).
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Selectionand
Selection andpeer-review
peer-review under
under responsibility
responsibility of
of Asociatia
the Emerging Markets
Grupul Roman Queries in Finance
de Cercetari and Business
in Finante local organization.
Corporatiste
Keywords: data analytics; preventing fraud; IT systems.

* Corresponding author. Tel.: +40-721-335-620.

E-mail address: rescuadrian@yahoo.com.

2212-5671 © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Selection and peer-review under responsibility of Asociatia Grupul Roman de Cercetari in Finante Corporatiste
doi:10.1016/S2212-5671(15)01485-9
1828 Adrian Bănărescu / Procedia Economics and Finance 32 (2015) 1827 – 1836

1. Introduction

Economic and financial crisis, which began in 2008 with the bankruptcy of Lehman Brothers in the USA,
unlike previous crises, has had a very rapid nationally and internationally spread, so that, in 2009 it established
its systemic nature, most countries being affected directly or indirectly by a strong recessionary phenomena,
imbalances and turbulence in real and nominal economy.
This has affected the industrialized world by lack of credit and falling property prices. Slowing economic
growth of developed countries produced effects on emerging economies, major EU trading partners, with
negative effects on exports. No doubt, the economic downturn will be felt by modern societies for years to come.
As you could easily find, these manifested differently from state to state, depending on the features related to
their debt exposure to the so-called "real estate boom", the ability to innovate and compete, resistance to fraud.
In this context, economic crime and fraud remains an intractable problem for global companies. According
to 2014 ACFE report, organizations lose 5% of revenues each year to fraud. If applied to the 2013 estimated
Gross World Product, this translates to a potential projected global fraud loss of nearly $3.7 trillion. Also, the
median duration (the amount of time from when the fraud commenced until it was detected) for the fraud cases
reported to us was 18 months. We believe that fraud mechanisms are more important than ever, being a key tool
for maintaining productivity growth conditions and sustained economic growth.

2. The current state of implementing data analysis software for preventing and detecting fraud

In the last few years, we have been witnessing a massive increase in the quantity of data (text, pictures, audio,
video etc.), both at global and economic level entities. This process is amplified by the entry of any entity
mentioned above into the virtual environment. Data comes from everywhere, from numerous and diverse
sources like contracts, customer interactions, call centers, social media, phones, emails, faxes, and others. The
trend is to use these data for the interest of the entity (conceiving strategies, opportunities identification,
goodwill development, preventing and detecting fraud etc.).
The use of data analysis processes and the software dedicated to these operations provide extensive and in-
depth analysis of the phenomena and processes of the informal economy, fraud and corruption, as the
information and communication technology becomes a sine qua non instrument of registered (formal) economy.
Although on the analytical market, there is a wide spectrum of specialized tools capable to support and enhance
the antifraud activity, unfortunately, the survey results indicate that managers are not taking advantage of them.
Forensic data analytics (FDA) tools are currently in use in the organizations, but there is much lower adoption
of more sophisticated FDA tools as depicted in Table 1. 65% of survey participants report the use of spreadsheet
tools such as Microsoft Excel and 43% report the use of database tools such as MS Access or MS SQL Server.
While these tools are important to every FDA program, they often focus on the matching, grouping, ordering,
joining or filtering of data that is primarily descriptive in nature.
 Adrian Bănărescu / Procedia Economics and Finance 32 (2015) 1827 – 1836 1829

Table. 1. Forensic data analytics tools use in the organizations

Forensic data Percent
Spreadsheet tools such as Microsoft Excel 65%
Database tools such as Microsoft Access or Microsoft SQL Server 43%
Continuous monitoring tools, which may include governance risk and 29%
compliance (GRC) tools (SAP, SAI Global, Oracle)
Text analytics tools or keyword searching 26%
Forensic analytics software (ACL, iDEA) 26%
Social media/web monitoring tools 21%
Visualization and reporting tools (Tableau, Spotfire, QlikView) 12%
Statistical analysis and data-mining packages (SPSS, SAS, R, Stata) 11%
Big data technologies (Hadoop, Map Reduce) 2%
Voice searching and analysis (Nexidia, NICE) 2%

Source:http://www.ey.com/Publication/vwLUAssets/EY-Global-Forensic-Data-Analytics-Survey-2014/$FILE/EY-Global-Forensic-Data-
Analytics-Survey-2014.pdf

According to the conclusion of the 2014 ACFE report, it should be noted that the authors identify the
proactive data monitoring/analysis as one of the most effective tool for antifraud control, in helping reduce fraud
losses and fraud scheme duration. Therefore, 34,8% of organizations affected by fraud used proactive data
monitoring/analysis as a control tool; organizations using proactive data monitoring/analysis faced with a near
59,7% reduction in median loss compared to those that did not; organizations using proactive data
monitoring/analysis experienced a 50% reduction in median duration of fraud scheme compared to those that
did not. The main barriers of implementing an analytical system are presented below in Table. 2.

Table 2. The greatest barrier in leveraging data analytics/CAATs when performing audit activities

Forensic data Percent

The requisite skills and knowledge do not currently reside within the internal 34% 21% 18% 14% 12%
audit department.
Lack of access to data analytics/CAATs tools 20% 23% 21% 23% 13%
Lack of audit methodology and approach supporting the use of data 12% 28% 35% 19% 6%
analytics/CAATs
Use of data analytics/CAATs is not perceived as a means to increase the 13% 18% 19% 34% 15%
efficiency of conducting audit activities.
Other 44% 10% 9% 5% 32%

Source: PricewaterhouseCoopers (2010) State of the internal audit profession study

By authors of 2012 AuditNet survey, among the relative important factors, regarding decisions on integrating
data analysis in the audit process, data quality was the most important factor, and the fraud detection situated at
the fifth level from thirteen.
Table 3. The relative importance of factors, regarding decisions on integrating data analysis in the audit process

Answer Options Extremel Important Not Very Not at all Rating Response
y Important Important Average Count
Important
Software cost 54 102 26 5 3,10 190
Technology capabilities of our staff 47 124 16 3 3,13 192
Training availability 42 125 20 2 3,10 191
1830 Adrian Bănărescu / Procedia Economics and Finance 32 (2015) 1827 – 1836

Training costs 50 114 20 5 3,11 191

Staff retention 46 94 37 9 2,95 191
Making the audit process more 95 82 10 1 3,44 190
efficient
Finding fraud 72 91 23 2 3,24 189
Audit Committee (Board support 36 71 56 10 2,77 190
Senior Management support 64 73 41 9 3,03 191
CAE support 79 71 16 4 3,32 188
Requires adjusting 19 83 71 13 2,58 192
our audit work programs
Getting access to the data 91 88 11 0 3,42 191
Data quality and reliability 11 66 11 0 3,53 190
1
Source: AuditNet 2012 Survey Report on Data Analysis Audit Software
In our opinion, there are all prerequisites for the data analysis to have a great potential for changing the way
that fraud detection is done. The use of analytical instruments may be deducted from specific indicators, like:
• the raise of fields based solely on data/information/intangible assets (a direct consequence of the ageing
process of knowledge society);
• new categories of analyst profession (intelligence analyst according to 2012 Classification of Occupations
in Romania);
• crystallization of diverse markets which offer software tools for data analysis (a direct consequence of the
objective laws of supply and demand);
• strategic acquisitions made by market giants in the field (I2 acquisition by IBM Corporation).

3. Possible solutions to improve data analysis processes for preventing and detecting fraud

Data analysis can be described as in-depth examination of the meaning and essential features of available
data, in order to identify significant information, using specific methods and techniques. It’s an interdisciplinary
domain which includes branches such as science computers (computer science), mathematical sciences,
statistical, economic, psychology, law and other cognitive sciences.
This careful examination of data identifies data gaps, strengths, weaknesses, dysfunction, vulnerabilities and
risk factors that may constitute threats and finally suggests guiding lines. Although on the field there are several
concepts of data analysis such as intelligence analysis, business analysis etc., they all have common
components. The differences depend on the scope, nature of the data, analytical products, practical utility and
applicability.
According to specialized literature, related to data analysis as system for prevention and detection of fraud,
can be identified over 24 types of analysis, some of them extremely complex, but among all of them, we can
identify two classical types of analysis: operational analysis and strategic analysis. The basic concepts of data
analysis are emphasized in the table below:
 Adrian Bănărescu / Procedia Economics and Finance 32 (2015) 1827 – 1836 1831

Table 4. Basic concepts of data analysis

Methods Strategic analysis Operational analysis

Techniques Risk analysis, Results analysis, Phenomenon Case analysis, Comparative case
analysis, Situational picture analysis, statisticalanalysis
analysis, SWOT Analysis Links analysis, Flow analysis, Event
PESTEL Analysis, Scenarios technique analysis
Analysis of activities, Financial
Analysis, Analysis phones, Risk
analysis, SWOT Analysis, Profile
Suspect Analysis, Geospatial Analysis,
Technology scenarios etc.
Procedures Graphical representation (histograms, relation maps, flow maps, maps of activities, of
events, geospatial maps), Space viewing, Three dimensional Viewing etc.
Instruments Mathematics, Statistics, Office Excel, Access, SAS, iDEA, GeoMedia Professional, GPS,
Map, ANB, iBase, Palantir, paper, pencil, etc.

Source: Cofan S. M. et al., 2014

Operational analysis can be successfully used on short term, by exploiting data and current information for
the purposes to comply with the current activities for detection of fraud with maximum efficiency. Data analysis,
in operational form, becomes a tool for improving workplace conditions, most manual activities are avoided,
reduced mental effort, especially for operations involving the processing of a significant amount of data. In daily
activities, the main role of operational analysis is to help antifraud manager to detect and combat illegal
activities, by examining (i) links between suspects, (ii) their characteristics (direct or/and indirect subordination
relations, positions in the hierarchy of the group, key positions that impact decision-making etc.), (iii) the
movement of goods, money or other valuables, (iv) way of communication (email, social networking), (v)
sequence of certain events, (vi) modus operandi etc. The success of such approach lies in the quality and variety
of data sources.
One of the purposes of using processes analysis of the data, in the context of the operational analysis, is to
complete the information gaps or to remove uncertainties and contradictions. Finally, the analytically product
must be submitted in a clear and concise form, containing at least one mode of illegal operation or a worthy
presumption of fraud Unlike operational analysis, the strategic analysis involves a macro-level approach of
preventing and detecting fraud issues. Considering that analytical tools are much more diverse, in case of
strategic analysis the amplitude of the activities is exacerbated. In this context will be studied threats,
vulnerabilities, risks, trends of evolution of fraud phenomena, the evolution of the market, demographic aspects
of fiscal policy, economic development or decline of the entities. It will scan both internal environment, with
vulnerabilities and institutional capacities and the external context with opportunities and potential threats.
The analyst will work with large volumes of data, it will use statistics, will describe phenomena and will
formulate explanations and predictions, in order to substantiate decisions for the highest level management.
Metaphorically speaking, research analyst shall look at the relevant object like “a hawk from the sky”, able to
distinguish rather color spots than the details (Cofan et al, 2014), (Comes, 2013). The strategic analysis offers a
macro overview related to fraud. Analytic tools represents piece of software which improve methods or multiply
their effects. If utilizing digital statistical tools, whether or not you have training in statistics, you can identify,
for example, unusual variation by looking to the shape of a curve.
The actual statistical tools have a highly intuitive interface and a powerful statistical processing engine. For a
long time, analysis techniques, such as statistics and exploration of data, were developed independently of the
visual techniques. With the new generation of visualization software, we can dive into massive data sets and
visually find new trends, patterns and threats that would take hours or days using conventional data mining
(Bresfelean et all, 2008 ).
1832 Adrian Bănărescu / Procedia Economics and Finance 32 (2015) 1827 – 1836

Data mining, as an analytic process, is designed to explore data, to extract information from data sets, in
order to discover patterns and relations. It can be defined as “the nontrivial extraction of implicit, previously
unknown, and potentially useful information from data” (Frawley et al., 1992, Bresfelean et al, 2007), or “the
science of extracting useful information from large data sets or databases” (Hand and Mannila, 2001). Analysis
of the data-text, known under the name of exploitation of data such as text or "text mining", refers to the process
of knowledge extraction from documents, because information can be mostly found in text format. Occurrence
of this form of analysis is associated with the moment when classical research methods became inefficient. The
indexing methods surpassed the issue mentioned above, the search algorithms contained rules in order to
accomplish a variety of analytic tasks, as categorizing documents, creating summaries, detecting relevance
between documents etc. Text data mining is considered to be one of the most important area in the database
system, which provides one of the most interesting and promising development in informatics industry
(Cre‫܊‬ulescu, 2011). Some authors (Ah-Hwee Tan, 2012) identifies and summarizes the capabilities of 11
(eleven) text data mining products, organized in two clusters. First cluster of products optimizes activities of
organization, viewing and navigation within documents, and the second one provides functions for text analysis,
in particular, the extraction, classification and summarization of information.
9
1
Summarization
Visualization
8
Sentiment 1
Analysis Text
Identification
7
2
Link Analysis
Text Mining

6 Text Anaytics 3
Predictive (c) predictiveanalyticstoday.com Text
modelling Categorization
5 4
Text
Search Acces
Clustering

Fig. 1. Text analytics process flow

Geospatial analysis - visual analysis is also important for understanding the relevance of the location where
events happened, to determine and discover patterns in fraud behavior. Analytical instruments allow
identification, exploration, indexing and processing data. The biggest challenge remains implementation of
automatic methods and tools. It is absolutely necessary to analyze possibilities for expansion and integration of
the new models and computerized systems to prevent and detect fraudulent actions and to support wittingly
managerial decisions. Successful implementation of an antifraud analytical system highly depends on the
manner of retrieving data from a variety of sources, considering that most of them have different formats. It is
recommended that the data collected should be interpreted in the same way, using the same techniques and the
same methodology, so that creation of data bases to be homogeneous. Although a system of fraud prevention
and detection could be expensive, the low degree of response to fraud and the inability to recover losses resulting
from internal fraud or abuse could have multiple consequences, difficult to quantify. The actual fraud prevention
and detection mechanism combines both human and technical factors. No matter how sophisticated technical
 Adrian Bănărescu / Procedia Economics and Finance 32 (2015) 1827 – 1836 1833

solutions may be, human factor still dictates the action mode as well as the exploitation of results. Introduction
of a computerized system of control would eliminate the problem of inefficiency of control levels, referred to
at points 1, 2 and 3 in Figure no. 2 below. These levels are associated with the unforeseen or repetitive inspection
of work, specific to the control activity and are characterized by a low percentage of early detection of fraud.
Through the leverage of automated controls and monitoring tools (level 5), it can be mitigated potential risks to
misconduct and fraud.

Fig.2. The efficiency of control levels

Source: adaptation of model http://www.acl.com/pdfs/White_Paper_AACM.pdf

Last but not least, integration of data analysis processes within the fraud detection system is an imperative
arising from scientific and technological development of allover companies, accompanied by a series of benefits
and limitations as follows.
(a) Benefits:
• get answers, in real time, to a series of questions regarding fraud issues;
• automatic data collection (predetermined flow);
• total and fast access to all data, through data indexing software (way of sorting a number of records on
multiple fields.);
• eliminates double records, errors, improving quality of data;
• high productivity vs. manual work;
• operating with incomplete and inaccurate data;
• generating a positive yield and fast return on investment;
• an increased rate for fraud detection;
• fast detection and recovery of consequences of fraud activity;
• creation of statistical analysis with high degree of accuracy;
• reducing fraudulent claims;
• increase the quality of analytical products.
(b) Limitations:
• like the other labor saving tools, fraud prevention and detection software do not come cheap;
• large part of data are not introduced in databases, not all text files being included in the final reports;
• the utilization of analytical tools don’t save time, just optimize it. The time gain like this, is used for further
research/analysis;
1834 Adrian Bănărescu / Procedia Economics and Finance 32 (2015) 1827 – 1836

• regardless of the software complexity, a human resource is always needed;

• efficient antifraud system involves high costs, so, many of economic entities prefer to create only classical
control structures;
• anti-fraud activities, based on software and hardware solutions, are recommended to be carried out and
coordinated by a group of specialists, with different experience, in order to cover various fields of activity;
• the absence of audit programs is a vulnerability and for safety reasons, the access to information must be
controlled in both ways, in-and-out;
• Due to the complexity of analytical research, the final product can be hard to assimilate, so, it’s
recommended to use descriptive part (interpretation of tables, graphs, values, metadata etc.).

4. Conclusion

Our intention is to encourage antifraud managers to use proactive data detection techniques in order to
improve fraud prevention and detection. There is not a toolkit which you can start a business fraud detection, is
not recommended to spend too much time selecting the perfect option. Just get started fighting fraud, use free
and payable software, a combination of statistical, data visualization, data mining, and filtering tools. The
processes of data analysis as a tool for preventing and detecting fraud can be used successfully in any field,
mainly in those of the database and the data are or may be easily converted into electronic format. For the fiscal,
banking, insurance and medical fraud existence of a structure is a sine qua non for the survival of business in
the current exacerbation of fraud, financial constraints and fierce competition. Although the software are not
cheap, as we have previously mentioned above, there is the possibility to maximize the benefits offered by the
Office package (Excel, Access) or ActiveData for Excel/Office. Creating a system to detect and prevent fraud
involves certain steps, which can be done gradually, depending on the priorities and the complexity of the
system, as further presented related to hardware components:
• define the intention of preventing and detecting fraud;
• create a special entity for this purpose;
• create an it infrastructure able to transpose the internal and external data into the virtual domain;
• ensuring a flow of creating and storage of data in electronic format;
• implementation of a monitoring system data to allow, where possible, real-time detection of irregularities
so as to avoid damage. The system should contain a number of templates (predefined models) built for
detecting fraud schemes. As architecture, it is recommended to be partially predefined, so as some modules
can be customize for the customer’s needs;
• creating a recovery system;
• develop an integrated data analysis (a nucleus together with the most detection methods: statistical,
relational etc.);
• creating a system able to generate intermediary and final reports, depending on the requirements of the
recipient.
Acknowledgements
This paper has been financially supported within the project entitled „SOCERT. Knowledge society,
dynamism through research”, contract number POSDRU/159/1.5/S/132406. This project is co-financed by
European Social Fund through Sectoral Operational Programme for Human Resources Development 2007-
2013. Investing in people!
 Adrian Bănărescu / Procedia Economics and Finance 32 (2015) 1827 – 1836 1835

References
ACL 2014, Fraud Detection Using Data Analytics in the Banking Industry. Discussion whitepaper.
Alexandru, D., I, 2010. Manual de Inteligen‫܊‬a Afacerilor, ISBN-13: 978-1446694251.
Argyriou, E.N.,Symvonis, A. 2012, Detecting periodicity in serial data through visualization. Advances in Visual Computing, vol. 7432,
pp. 295-304.
Association of Certified Fraud Examiners. 2014. Report to the Nation on Occupational Fraud and Abuse,
AuditNet 2012 Survey Report on Data Analysis Audit Software.
Bresfelean, Vasile Paul, Mihaela Bresfelean, Nicolae Ghisoiu, and Calin-Adrian Comes. 2007. "Data Mining Clustering Techniques in
Academia." In ICEIS (2), pp. 407-410.
Bresfelean, V. P., Bresfelean, M., Ghisoiu, N., & Comes, C. A. 2008. Determining students’ academic failure profile founded on data mining
methods. In Information Technology Interfaces, IEEE, pp. 317-322.
Burge, P., Shawe-Taylor, J. 2001, An Unsupervised Neural, Network Approach to Profiling the Behaviour of Mobile Phone, Users for Use
in Fraud Detection. Journal of Parallel and Distributed Computing 61: 915–925.
Chersan, I.C., Carp, M., Mironiuc, M., 2013, Data mining – o provocare pentru auditorii financiari, revista Audit Financiar, anul XI, nr.
106, 10/2013, pag. 57- 64.
Cofan S.M., Ivan, L., Dogaru V., Cios A., Savin M. 2014. Analiza Informa‫܊‬iilor. Manual, ed. Ministerului Afacerilor Interne, ISBN 978-
973-745-129-3.
Comes, C.A. 2013. Stored Procedure Migration In ERP Systems, LAP Lambert Academic Publishing, Saarbrucken, Germany.
Comes C.A, Bresfelean V. P., 2013, Social Networking Analysis of ReCADD with Petri Nets, Cambridge Scholars Publishing, Newcastle
upon Tyne, UK, pp. 25-33.
Cox, K., Eick, S., Wills, G. 1997. Visual Data Mining: Recognizing Telephone Calling Fraud. Data Mining and Knowledge Discovery 1:
225–231.
Cretulescu, R. G., 2011, Contributions to the design of classification systems of the documents. Legislation, Sibiu.
D.Hand, H. Mannila, P. Smyth: Principles of Data Mining. MIT Press, 2001
Frawley W., Piatetsky-Shapiro, G., Matheus, C., 1992, Knowledge Discovery in Databases: An Overview. AI Magazine, p. 213-228
Global fraud study, 2012, Report to the nations on occupational fraud and abuse
Green, B., Choi, J. 1997, Assessing the Risk of Management Fraud through Neural Network Technology. Auditing 161: 14–28.
http://aitfis.com/image/AIT_1811ArticleWeb.pdf
http://linkanalysisnow.com/2010/11/using-visual-analysis-to-detect-fraud.html
http://www.acfe.com/rttn/docs/2014-report-to-nations.pdf.
http://www.acl.com/pdfs/DP_Fraud_detection_BANKING.pdf.
http://www.auditnet.org/publications--2/auditnet-surveys
http://www.ey.com/Publication/vwLUAssets/EY-Global-Forensic-Data-Analytics-Survey-2014/$FILE/EY-Global-Forensic-Data-
Analytics-Survey-2014.pdf.
http://www.tide.org.tr/uploads/DetectingFraudBook.pdf.
http://www-01.ibm.com/common/ssi/cgi-bin/ssialias?infotype=SA&subtype=WH&htmlfid=YTW03084USEN
IBM, Step-by-step data mining guide,
Kent, K., Chevalier, S., Grance, T., Dang H. 2006, Guide to Integrating Forensic Techniques into Incident Response,
http://csrc.nist.gov/publications/nistpubs/800-86/SP800-86.pdf.
Lanza, R.B. 2003, Proactively Detecting Occupational Fraud Using Computer Audit Reports,
Mantone, P. S 2013, Using Analytics to Detect Possible Fraud: Tools and Techniques, ISBN-13: 978-1118585627
McMahon, R.J. 2001, Practical Handbook for Private Investigators, CRC Press LLC, ISBN 0-8493-0290-0.
Michalski, R. S., Bratko, I., Kubat, M. 1998. Machine Learning and Data Mining – Methods and Applications. John Wiley & Sons Ltd.
Millar, P. 2009, Detecting and Preventing Fraud with Data Analytics. http://www.acl.com/pdfs/eBook_fraud.pdf
Murad, U., Pinkas, G. 1999, Unsupervised Profiling for Identifying Superimposed Fraud. Proceedings of PKDD'99.
Nelson, J. Chen. 2011, Beyond the Next Generation Technology for Financial Crimes and Asset Forfeiture Investigations. The Eighteen
Eleven: Professional Journal of the Federal Law Enforcement Officer Association, no. 1, pp. 6-7,
Nigrini, M. 2011, Forensic Analytics: Methods and Techniques for Forensic Accounting Investigations. Hoboken, NJ: John Wiley & Sons
Inc. ISBN 978-0-470-89046-2.
Ni‫܊‬u, I., 2012, Analiza de intelligence. O abordare din perspectiva teoriilor schimbării, Editura Rao, ISBN: 978-606-609-373-6.
Oppel, A. 2011, Modelarea datelor. Ghidul începătorului, Editura: Rosetti Educa‫܊‬ional,
Palshikar, G.K., 2002, The Hidden Truth – Frauds and Their Control: A Critical Application for Business Intelligence, Intelligent Enterprise,
vol. 5, no. 9, pp. 46–51.
Phua, C., Lee, V., Smith-Miles, K. and Gayler, R. 2005, A Comprehensive Survey of Data Mining-based Fraud Detection Research. Clayton
School of Information Technology, Monash University.
PricewaterhouseCoopers LLP 2009. "2009 Global Economic Crime Survey". Retrieved June 29, 2011.
Robu, I.B., Robu, M.A., 2013, Proceduri de audit pentru estimarea riscului de fraudă bazate pe indici de detectare a manipulărilor contabile,
revista Audit Financiar, anul XI, nr. 106, 10/2013, pag. 3-16.
1836 Adrian Bănărescu / Procedia Economics and Finance 32 (2015) 1827 – 1836

Smith, S., Mueller, J. 2012, Data Analysis Challenges: Try Proven Strategies for More Success. Association of Healthcare Internal Auditors.
https://www.audimation.com/pdfs/ahia-data-analysis-challenges.pdf.
Spann , D. D. 2013, Fraud Analytics: Strategies and Methods for Detection and Prevention, ISBN-13: 978-1118230688.
Westphal, C., 2009, Data Mining For Intelligence, Fraud, & Criminal Detection Advanced. Analytics & Information Sharing Technologies.
CRC Press Taylor & Francis Group. 978-1-4200-6723-1.
Young, M.R., 2014, Financial Fraud Prevention and Detection. Governance and Effective Practices, John Wiley & Sons, Inc., Hoboken,
New Jersey, ISBN 9781118617632.